startup-launch-kit
Pass
Audited by Gen Agent Trust Hub on Jun 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill operates as a high-level orchestrator for a set of known startup-related skills. It focuses on sequencing and state management without performing unauthorized actions.
- [COMMAND_EXECUTION]: The skill uses the
Skill()tool to invoke subordinate skills (brand-workshop,validation-canvas, etc.) in a defined sequence. This is the intended functionality for an orchestrator and is performed within the agent's controlled environment. - [DATA_EXFILTRATION]: In 'existing-project' mode, the skill triggers
startup-auditto read local codebase files (e.g., dependency manifests, routes) to infer project details. This exposure of project data is used locally to pre-populate a validation canvas and requires explicit user confirmation/correction, mitigating risks of automated processing of untrusted data. - [PROMPT_INJECTION]: The skill processes untrusted data from the user's codebase (Category 8: Indirect Prompt Injection). It implements a 'human-in-the-loop' verification step where the user must review and confirm machine-inferred content, which serves as a significant safety boundary.
Audit Metadata